Cars, Air Pollution and Low Emission Zones in Germany

1

Cars, Air Pollution and Low Emission Zones in Germany1

Lisa Perry and Hendrik Wolff

VERY PRELIMINARY -- COMMENTS ARE WELCOME

PLEASE DO NOT CITE OR DISTRIBUTE

VERSION June 21, 2009:

University of Washington

Abstract:

Low Emission Zones (LEZ) are areas in which vehicular access is restricted to only vehicles that

emit low levels of air pollutants. Recently, LEZs have become a popular policy worldwide, and in

Germany LEZ are increasingly adopted to comply with the 2005 EU air pollution legislation. We

study the effect of LEZs in two ways. First, we assess whether pollution levels do actually

decrease. Second, we analyze the spatial substitution effects in new vehicle purchases and

retrofits of high emission cars due to the LEZ regulation. We find that LEZs significantly

decrease air pollution in urban centers. We, however, also find that outside of the LEZ pollution

increases likely because the higher polluting vehicles are forced to drive longer routes around

the LEZ. Moreover, we find that German vehicle owners substantially increase the adoption of

cleaner technologies the closer they live to an LEZ. In summary, if marginal damages are

convex, the overall effect of the LEZ program on air pollution could be positive.

Keywords: Air Pollution, Low Emission Zones, PM10

JEL: Q20

1 We like to thank the Umweltbundesamt of Germany to providing the air pollution data and the Kraftfahrtbundesamt of Germany for help discussing the vehicle emission categories. We are grateful for the research grant provided by the Royalty Research Fund of the University of Washington.

2

1. Introduction

Increased health concerns about air pollution have led many countries to tighten traffic-

related policies. Legislators have focused on PM10, defined as the class of particles smaller than

10 micrometers (μm), as PM10 has the ability to enter directly into lungs and partially into the

bloodstream. It is estimated that PM10 causes 348,000 premature deaths per year in the

European Union (EU), while Ozone—the second most deadly air pollutant—causes about 21,000

premature deaths (Watkiss et al., 2005).

The EU commission has taken these results very seriously and enacted the 2005 EU Clean

Air directive, which marks an unprecedented attempt to mandate low levels of PM10. When

cities violate the maximum allowable limits on PM10 (by exceeding 50μg/m³ per day on more

than 35 days or averaging more than 40μg/m³ within a calendar year), mayors and local

governments have to develop so called clean air action plans. The contents of these actions plans

are determined by the city council but they typically consist of building ring roads and enhancing

public transportation. The most drastic measure, however, is the implementation of a Low

Emission Zone (LEZ), which defines an urban area where higher-polluting vehicles are - banned

from driving.

In Germany, to deal with the large number of cities exceeding the EU PM10 threshold, the

government has categorized all vehicles into four mutually exclusive classes of PM10 emissions.

All automobiles (all 46 million German vehicles and all visiting foreign cars and trucks) are

required to display a colored windshield sticker indicating which class it belongs to—green,

yellow, red or no sticker in the order of the lowest to the highest polluting category. Each city

can choose which sticker cars it permits into the LEZ. Currently Germany has 32 LEZs which

vary in size from just a couple of blocks (particularly in small towns) up to 207 square

kilometers in the case of Stuttgart, where the entire town and large portion of the suburbs are

affected (see map in Appendix E). While LEZs have become a popular quick fix for local

governments to meet the EU PM10 legislation, they have been very controversial because of the

significant costs imposed on drivers and business owners, as well as bus and truck companies for

whom upgrading fleets to the appropriate sticker is very expensive.2

2 For smaller vehicles, the conversion to the next higher sticker costs 800 to 2500 U.S. dollar. For larger vehicles and trucks the conversion costs 10,000 to 22,000 dollar. For some vehicles the conversion is technologically infeasible.

3

Germany is not alone in limiting vehicle use. In fact, driving restrictions have been used for

decades in some of the world’s most polluted cities. In 1989 Mexico City introduced the Hoy

No Circula (HNC) policy which prohibits driving between 5am and 10pm one weekday per week

based on the last digit of its license plate. Similar restrictions have been implemented in Athens

(1982), Bogota (1998), Santiago (1986) and São Paolo (1997), San Jose (2005), La Paz (2003),

all of Honduras (2008), and Beijing (2008). Other forms of driving restrictions include partial

and total bans (Italy, Athens, Amsterdam, Barcelona, and Tokyo); traffic cell architecture, such

that vehicles can drive within the cell but must take a circumferential ring road to get between

traffic cells (Goddard 1997; Vuchic 1999); traffic bans on days when air pollution exceeds

certain thresholds, such as in Milan and other Northern Italian cities; and congestion charging in

combination with an LEZ (such as in London where low emission vehicles do not pay the

congestion charge but all other automobiles pay 8 British pounds (=14 U.S. dollar).3

Despite the widespread use of driving restrictions, the related empirical literature is sparse.

In a recent study, Davis (2008) analyzes the effect of Mexico City’s HNC policy on air quality.

While he finds no change in weekday pollution levels, pollution actually increased on the

weekends and weekday late nights as drivers substituted towards driving when the HNC was not

in effect. Davis shows this ineffectiveness is due to a surprising behavioral response: drivers

circumvented the restriction by buying older, more polluting vehicles to be second cars with

different license plates4. Davis also shows that the HNC is a high-cost solution—with social

costs exceeding $300 million per year—given its negligible effect on air quality.

While the counterproductive results in Mexico City were due to the particular design of the

HNC5, the German LEZ program may be more successful because it includes a differentiation by

3 Price-based policies that aim to limit congestion and emissions include road pricing and congestion fees. Singapore (1975), London (2003) and Stockholm (2006) all charge fees for driving into the city center. While New York City's plan to introduce congestion fees stalled in the legislature, San Francisco is currently debating plans to implement a six dollar fee to drive through downtown. Milan has combined congestion pricing and LEZs with its Ecopass program, which charges fees to enter the downtown area based on emissions-level of the car. Even with increasing adoption of such price-based policies, command and control driving restrictions are more often adopted because these policies are easier to implement politically, technologically more feasible, and relatively less expensive to enforce (Levinson and Shetty, 1992; Davis, 2008). 4 Drivers also increased their use of taxis, which were some or the most polluting vehicles in Mexico City when the policy was first enacted. 5 The HNC has been modified to include an exhaust monitoring program, Verifcación, such that each car is affixed with a sticker indicating their class of emissions and the cleanest of the classes are exempt from the HNC.

4

emission level, creating incentives to adopt cleaner technologies.6 Even with this incentive,

though, such command and control policies may prove inefficient since they lack market-based

incentives that equalize the marginal cost of abatement across polluters. LEZs may also cause

unintended consequences such as increased driving outside of the LEZ, especially by the more

polluting vehicles that cannot enter the LEZ.

Therefore, whether LEZs are effective is an empirical question. We think, that this is also a

very urgent question, since in 2010 the EU PM10 standards will be drastically tightened and as a

consequence many more cities will have to aggressively implement air control strategies7.

To this end, the first task of this paper is to estimate the causal effect of LEZs on PM10

levels using panel data of hourly PM10 levels and weather conditions across Germany from 2005

through 2008. Both the pre-regulation PM10 levels and the staggered nature of LEZ

implementations produce rich identification of the zones’ treatment effects. Using difference-in

differences regressions we analyze whether being in non-attainment status—and thus

implementing an action plan—has had an effect on PM10 levels. We then look at whether this

effect is different for cities whose action plans did and did not include LEZs.

One important argument in favor of the four-tier PM10 categorization is that it leads to a

more rapid adoption of cleaner technologies since even vehicle owners who do not typically

drive into an LEZ may want to keep the option value of free passage. The next task of this paper

therefore is to study changes in the composition of the vehicle fleet. Using a unique panel

dataset that provides the emission category and registration location of each privately and

commercially owned German vehicle from 2006 to 2009, we analyze the spatial substitution in

vehicles’ emission categories due to the LEZ regulation.

In our preliminary analysis we find that while action plans alone have no significant effect on

lowering PM10 levels, LEZs decrease PM10 by four to seven percent within the area of the LEZ.

We, find that at background stations off of major roads, however, pollution increases. This is

likely because higher polluting vehicles are forced to drive longer routes around the LEZ.

6 Evidence that a small percentage of high-emission vehicles contribute the bulk of pollution leads Roson and Small (1998) to argue that targeting dirty vehicles may be the most effective way to decrease emissions. For example, Small and Kazimi (1995) find heavy-duty diesel trucks have social costs per mile ten times higher than gasoline vehicles. 7 We estimate that based on the 2010 EU PM10 standards, currently 253 cities are in violation. The 2010 standard allows for seven (7) days only to exceed the 50μg/m³ per day average (instead of the current rule of 35 exceedance days. In addition the yearly average will be set to 20μg/m³.

5

Moreover, we find that German vehicle owners substantially increase the adoption of cleaner

technologies the closer they live to an LEZ. In summary, if marginal damages are convex, the

overall effect of the LEZ program on air pollution could be positive.

This papers proceeds as follows. Section two details the EU PM10 regulation and the

implementation of LEZs. We describe our data in section three and discuss the empirical

strategy in section four. Section five presents econometric results of the causal impact of the

LEZs on PM10 levels. Section six discusses the spatial substitution effects of high to low

emission cars and we conclude with policy recommendations in section seven.

2. Background

2.1. Air Pollution regulation in Europe

In response to concerns about the health effects of PM10,8 the 2005 EU Clean Air directive9

introduced EU-wide limits on ambient PM10 such that: (a) the daily average does not exceed

50μg/m³ on more than 35 days annually and (b) the yearly average must not exceed 40μg/m³. In

Germany, the 35 day limit has been particularly troublesome and has been exceeded by 81

cities.10 Starting January 1, 2010, however, these thresholds will be drastically tightened to a

yearly average of 20µg/m³ and a maximum of seven days exceeding 50µg/m³. We estimate that

with these stricter 2010 limits, 285 German cities would be in nonattainment based on the

emissions from 2005-2008.11 The policy changes and implications are summarized in Table 1.

Compared to other pollutants (i.e. NOx, SO2), the limits on PM10 are by far the most often

violated; As of 2007, 70 percent of EU cities greater than population 250,000 had violated the

limits at some point, and, as of 2006, all EU member countries except Ireland and Luxembourg

had cities in violation of the PM10 threshold, mostly due to traffic in urban city centers (Europa

Press Release, 2008).

8 PM10 have been linked to serious cardiopulmonary diseases, acute respiratory infection, trachea, bronchus and lung cancers. (EPA, 2004). Worldwide, about 6.4 million years of healthy life are lost due to long-term exposure to ambient PM10 (Cohen et al., 2005). 9 See the European legislation on air quality 1999/30/EC and 96/62/EC. 10 No German city violated the standard based on the 40μg/m³ annual limit that did not also violate the exceedance day limit. 11 Both of the 2010 standards are even violated by the national average city in each year since 2005, which ranges from 12 to 27 days exceeding the 50µg/m³ limit (instead of 7 days) with an annual average above the 20μg/m³ limit. See Table 2 for details.

6

Table 1: EU PM10 limits

Panel A: European Union PM10 pollution thresholds Phase 1 Phase 2

since 1 January 2005 starting 1 January 2010 Yearly average limit 40 µg/m³ 20 µg/m³

Daily average (24-hour) limit 50 µg/m³ 50 µg/m³ Allowed number of exceedences per year 35 7 Numbers of German cities violating the standard

81 285*

Panel B: Germany violations of PM10 limits 2005 2006 2007 2008

National average PM10 [µg/m³] 24.4 (5.2)

26.2 (5.5)

23.1 (5.3)

21.2 (4.9)

Mean number of days** above 50 µg/m³ 19.6 26.8 16.2 11.6

(20.9) (21.1) (15.8) (12.9) Cities in violation of 2005 standard 36 65 31 18 Cities in violation of 2010 standard 226 246 200 134 *The calculation of the expected number of cities violating the 2010 standard is based on the number of cities that would have violated the standard between 2005 and 2008 either because of exceedance days or high annual averages **: Average of the highest exceeding station per city; Standard deviations in parentheses

When any air pollution station exceeds the EU PM10 limit, the city must immediately

develop an “action plan” and in order to enforce the legislation, the European Court of Justice

(ECJ) can impose financial penalties. In fact, in January 2009 the EC initiated infringement

proceedings against 10 EU countries that have not attained the EU PM10 limit12. Moreover, EU

citizens are entitled by law to demand action plans from local authorities.13

12These 10 countries are Cyprus, Estonia, Germany, Italy, Poland, Portugal, Slovenia, Spain, Sweden and the U.K.. Given the difficulty many countries have had meeting the PM10 limits, in June 2008 the EC began allowing countries to request an extension for meeting PM10 limits from January 2005 to January 2010. These extensions are only granted when a country has taken steps to combat PM10 at the local, regional and national level, such that exceedances are unavoidable due to “site-specific dispersion characteristics, adverse climatic conditions or transboundary contributions.” “ (Europa Press Release 2009) 13 Following an important recent ruling in July 2008 from the European Court of Justice, EU citizens can force their local councils to rapidly tackle air pollution. The EU’s court was asked to judge the case of Dieter Janecek, a resident of Munich, who said that under the EU Air Quality Directive, the city of Munich is obliged to take action to stop pollution exceeding the specified target. The judges reconfirmed with prior statements that concluded that EU citizens are entitled to demand air quality action plans from local authorities if the EU limits are exceeded. The LEZ of Munich was then implemented in October of 2008.

7

2.2. Low Emission Zones in Germany

Road transport is the primary source of ambient PM10 in most European urban areas

(Krzyzanowski, 2005).14 Given this primacy of vehicle-based PM10, action plans (AP) try to

constrain emissions through expanding public transportation, utilizing ring roads or improving

traffic flow. However, implementing an LEZ has emerged as an important—and controversial—

element of the German action plans. German cities began instituting LEZs starting January 1,

2008. The LEZs mostly cover city centers, but vary in size. In Berlin, for example, the LEZ

covers 88 square kilometers (km2), populated by 1.1 million people. Munich’s LEZ covers 44

km2 with 431,000 inhabitants and Frankfurt’s LEZ spans 110 km2. The largest LEZs is in

Stuttgart, which covers 207 km² with 590,000 habitants, compared to nearby smaller LEZs like

Illsfeld with 2.5 km2 and 4,000 habitants or Pleidelsheim with 7km2 and 7,000 habitants. Figure

1 below shows a map of current and planned LEZs and Appendix B for a list of characteristics of

all LEZs.

Each German car—as well as each visiting foreigner—that wants to enter an LEZ must

display a colored windshield sticker based on EU-wide emissions categories. There are four

PM10 classes. The highest emitting vehicles obtain no sticker (and hence cannot enter any LEZ),

while red, yellow and green stickers are given to progressively ‘cleaner’ cars. In some cases

vehicles can improve one class by retrofitting the engine or diesel particulate filter. As Table 2

shows, there are two pollutant classes for gasoline-powered vehicles (Green and No Sticker) and

diesel vehicles can attain any of the four pollution classes.

The implementation date and the types of cars restricted by an LEZ vary across German

cities. In Berlin, for example, all vehicles with a red sticker and "cleaner" (yellow and green)

have been allowed into the LEZ since January 2008, while access will be restricted to green

stickers only starting January 1, 2010. The LEZ of Dortmund (Brackler Strasse), on the other

hand, has only permitted yellow and green sticker cars since beginning in January 2008. Of the

23 LEZs implemented in 2008, four began in January, eight began in March, one began in July

and the rest began in October. A summary of all the current LEZs can be seen below in Table 3.

14 Road transport is also largely responsible for all NOX, CO, benzene and black smoke emissions. Although historically these toxins have been debated, the magnitude in terms of threshold violations and health impacts (premature deaths) is substantially higher for PM10.

8

Figure 1: Current (in red) and Future (in yellow) German LEZs

Table 2: German vehicle stickers

Sticker categories

No sticker Red Yellow Green

Requirement for diesel vehicles

Euro 1 or worse

Euro 2 or Euro 1 with particle filter

Euro 3 or Euro 2 with particle filter

Euro 4 or Euro 3 with particle figure

Requirement of gasoline vehicles

Without 3-way catalytic

converter

Euro 1 with regulated catalytic converter or better

Source: Umweltbundesamt

9

The fine for violating the LEZ is 40 Euros plus one driver’s license penalty point15. There

are exceptions that allow certain emergency and other work related vehicles to enter the LEZs

without a sticker. These include agricultural and forestry tractors; ambulances and doctor’s cars;

vehicles driven by or carrying persons with serious mobility impairments; police, fire brigades,

Bundeswehr and NATO vehicles.

Table 3: German LEZ restrictions 2008 to 2012

15 There is a staggered system of consequences due to penalty points, ending in loss of driver license with 18 points.

City J F MA MJ J A S ON D J F MA MJ J A S ON D J F MA MJ J A S ON D J F MA MJ J A S ON D J F MA MJ J A S ON D

BerlinBochumBottropDortmundDortmund (Brackeler Straße)DuisburgEssenFrankfurt am MainGelsenkirchenHannoverIlsfeldKölnLeonbergLudwigsburgMannheimMünchenOberhausenPleidelsheimRecklinghausenReutlingenSchwäbisch GmündStuttgartTübingenBremenDüsseldorfHeilbronnHerrenbergKarlsruheMühlackerPforzheimUlmWuppertal

2008 2009 2010 2011 2012

10

3. Data

We collected a panel of German air quality readings from 2005 through October of 2008

from the Federal Environment Agency of Germany, the Umweltbundesamt (UBA). This data set

includes a combination of half-hourly, hourly or daily readings of PM10, for 1285 stations in 388

cities. All stations are characterized by the UBA as being traffic, background or industrial

stations based on their location. This data is merged with weather data from the German national

weather service, Deutscher Wetterdienst. We obtained hourly weather readings for 34 stations

and daily reading for 74 stations. Because the air quality and weather monitoring stations are not

in the same location, we use the geographic coordinates to match each air quality station with the

closest air quality station. We use the PM10 readings from only stations that have a weather

station within 50 kilometers distance that are no more than 300 meters higher or lower in

altitude. The primary weather variables are summarized in Table 4 below. Using the procedures

detailed in Appendix A for cleaning data, calculating daily weather and PM10 readings and

handling missing values, we end up with complete PM10 and matched weather data for 185

stations covering 122 cities.

Table 4: Summary of weather data Weather Variables Unit Mean St. Dev. Min Max

Daily average temperature 1C 9.6 7.7 -27 31

Daily min temperature 1C 5.6 6.9 -29.7 24.4

Daily max temperature 1C 13.8 8.9 -23.1 40.2

Daily avg. vapor pressure 1 hpa 9.9 4.1 0.2 26.9

Daily average air pressure 1 hpa 981 49 679 1047

Daily avg. relative humidity % 78.1 12.9 7 101

Daily avg. wind speed 1 m/s 2.6 1.1 0 10

Daily max wind speed 1 m/s 10.9 4.9 1.3 64.8

Daily avg. cloud cover Tenths 7.1 1.3 0 9

Sun in day 1 hour 4.8 4.4 0 16.7

Precipitation during day 1 mm 2.1 4.7 0 158

Snow depth cm 4.1 28.6 0 550

New snow depth cm 0.2 1.9 0 150

Figure 2 illustrates the classifications of cities that we use in our analysis. First, we divide

stations into 2 categories, ‘attainment cities’ that do not violate the PM10 limit (and thus do not

11

need to develop an action plan) and ‘non-attainment cities’ that develop an action plan. Next,

among the non-attainment cities we differentiate between ‘action plan only’ (APO) cities, whose

action plan do not include an LEZ, and ‘LEZ cities’ that do include an LEZ in their action plan.

Figure 2: Main structure of treatment and control cities.

An examination of average PM10 levels shown in Figure 3 below shows that pollution

levels have been decreasing since 2006 for attainment cities and non-attainment cities, including

both action plan only cities, and LEZ cities. While the daily averages do not appear to have

decreased by more in LEZ cities than the action plan only cities, there may be effect once we

control for covariates such as station type, weather, city characteristics, holidays and others.

Finally, our data set is complemented by covariates that control for state-level school breaks

and legal holidays, both obtained by Johannsen (2009) as well as city-level 2006 population data

obtained by the Federal Statistical Office Germany Genesis database16.

16 Surprisingly, many smaller German cities are not included in the Genesis population file. For these cities we obtained the population estimates by internet search, e.g. from local city websites.

Attainment cities (C) Non-attainment cities (AP)

LEZ cities Action plan only cities (APO)

12

1015

2025

3035

4045

5055

60A

vera

ge P

M10

2005m1 2005m7 2006m1 2006m7 2007m1 2007m7 2008m1 2008m7Year

Attainment cities Non-attainment cities

Average of annual daily PM10 averages at per city

1015

2025

3035

4045

5055

60A

vera

ge P

M10

2005m1 2005m7 2006m1 2006m7 2007m1 2007m7 2008m1 2008m7Year


Traffic stationsAverage of annual daily PM10 averages at per city

1015

2025

3035

4045

5055

60A

vera

ge P

M10

2005m1 2005m7 2006m1 2006m7 2007m1 2007m7 2008m1 2008m7Year


Background stationsAverage of annual daily PM10 averages at per city

Figure 3: Average of city annual PM10 levels

Note: Plots of average of city averages per year, with 95 percent confidence intervals

13

4. Empirical Identification Strategy

4.1. Difference-in-Differences Approach

Our primary analysis uses two main difference-in-differences (DD) specifications (see Meyer

1995; Bertrand et al., 2004). First, at a national level, we match cities based on PM10 levels in

2005, before the action plans came into effect. To determine the effect being found in violation

of the PM10 limits, we compare PM10 levels in the non-attainment cities with action plans (AP)

to those of the control attainment cities (C). To differentiate the effect of LEZs, we then separate

the LEZ cities from the action plan only cities (APO).

Second, our next identification strategy takes advantage of the staggered nature of the LEZ

introduction, comparing the earliest LEZ cities to cities whose LEZ has not yet come into effect

(FUTURE LEZ). Because of the heterogeneity across states in term of weather, geography and

regional policies, for this analysis we look at changes in PM10 at the major LEZs one-by-one,

comparing them to nearby future LEZ cities. Finally, within the LEZ cities, we compare PM10

levels at stations that are located within the LEZ to those stations outside of the LEZ. Figure 4

summarizes the different treatment categories of cities and stations.

Figure 4: Classification of cities treatment status

To indentify the effect of an action plan (AP) on PM10, we estimate Equation (1), where k

indexes city, i indexes station and t indexes time:

1 ln , , , ΨΧ , , , ,

Attainment cities (C)

Non‐attainment cities (AP)

LEZ cities

'First' LEZs

(before Oct 2008)

(LEZ)

Stations within LEZ

Stations outside LEZ

'Late' LEZs

(after Oct 2008)

(FUTURE LEZ)Action plan only

cities (APO)

14

The dependent variable , , is the average daily PM10 reading for each station. is

an indicator variable for whether a city is in non-attainment of the PM10 limit and

is an indicator for time periods after the assignment of non-attainment status and implementation

of an action plan. , is an indicator variable that

equals one for non-attainment cities after being found in non-attainment and implementing an

action plan, such that measures the treatment effect of the action plan. Χ , , includes station-,

city- and time-specific covariates such as weather variables17, holidays, station type and city

population. Because location of the air quality station has a large impact on pollution readings,

we include station fixed-effects in all models. Identification comes from the assumption that,

after controlling for changes in observables such as weather and time-invariant unobserved

heterogeneity between cities, PM10 levels would change the same over the time frame studied in

treatment and control cities in the absence of the PM10 regulation.

In our second specification, we separate the treatment of action plans that do and do not

include LEZs. To isolate the effect of LEZ’s we estimate (2), where is an indicator for

being an LEZ city:

2 ln , , , ΨΧ , , , ,

Because there are multiple time periods in which LEZs are introduced, the time indicator

variables, , are a vector of time periods (Imbens and Wooldridge, 2008). We include

year-month fixed effects for this vector, which inherently includes a dummy for the introduction

date of all the LEZs18. , is an indicator variable that equals one for cities having

LEZs after the introduction of its LEZ, such that measures the treatment effect of the LEZ.

When we compare LEZ cities to cities with LEZs planned but not in effect during our sample

(FUTURE LEZ), we use the same specification as in (2), with the FUTURE LEZ cities as the

control group.

17 Weather variables include daily values of temperature, temperature squared, maximum daily temperature, minimum daily temperature, 1-day lag temperature and maximum temperature, relative humidity, relative humidity squared, 1-day lag relative humidity, wind velocity interacted with whether it rained that day, maximum daily wind velocity, 1-day lag wind velocity, visibility, precipitation, precipitation squared, days without precipitation, temperature interacted with precipitation, temperature interacted with relative humidity, temperature interacted with wind, air pressure, 1-day lag air pressure. 18 There is one LEZ that started on 1/12/2008, but this was only one street in Dortmund.

15

Finally, we are interested in to compare traffic stations that are located strictly within an LEZ

to the traffic stations outside of the LEZ but in the same city. Analysis of (3) mirrors (2), except

that the indicator for having an LEZ, , and the treatment indicator,

, , equal unity only for stations within the borders of a city’s LEZ. We use this

specification to study the effect of the zones within LEZ cities, by comparing the station within

LEZs to other stations in the same city outside the LEZ.

3 ln , , , ΨΧ , , , ,

In all analyses, standard errors are clustered at the city-week level. This allows us to correct

for the serial correlation in PM10 levels and any heteroskedacity across stations within a city

(see Bertrand et al., 2004).

4.2 Matching Stations and Cities

Our identification relies on matching treatment and control cities based on a similar PM10

level prior to the implementation of the clean air action plans. Specifically, we match cities on

the annual daily average of the cities' highest-polluting station19 in 2005. The year 2005 is used

since this is the last year that PM10 levels were not affected by action plans or LEZs; at the end

of this year, cities in non-attainment of the regulation would have had to institute action plans.

Table 5 shows the similarity in 2005 highest-average PM10 levels for the cities used in this

analysis. The complete list of cities in Appendix C shows that there are 80 cities that have 2005

highest-station PM10 averages in the range of 25 to 33.5 µg/m³. We do not use cities with PM10

levels below and above the 25 to 33.5 range, since at those ranges there do not exist appropriate

pairings of treatment and control cities based on the 2005 PM10 level. Cities found in non-

attainment (based on exceedance days) in 2005 are mixed throughout this range, meaning the

attainment cities and the non-attainment cities had relatively similar PM10 characteristics in the

base year of 2005. We assume, then, that within this group of cities, the 35 day threshold makes

the designation of non-attainment status and subsequent development of action plans exogenous.

Of these cities, 31 are cities that have never violated the PM10 limits or have had an action plan,

19 We use the station with the highest 2005 average PM10 reading to match cities since the exceedance of the PM10 regulation is determined by the highest-polluting stations, not the averages across all stations, and thus it is the highest reading within the city that is most important in determining PM10 levels.

16

which hence serve as our controls C. Nine cities violated the limit in 2005 or 2006 and have an

action plan but no LEZ, and 4 cities violated the limit in 2005 or 2006 and have implemented an

LEZ before October 2008. There are also four cities that implement an LEZ in October 2008,

six cities that have LEZs beginning in 2009 and 2010, and 26 cities that are considering

implementing an LEZ as part of their action plan. Since we only have data through October

2008, we count these as action plan only (APO) rather than LEZ cities. We exclude 17 of these

cities who either developed an action plan despite never violating the PM10 limit20 or didn’t

violate the limit until 2007 (but violated the limit i 2007 or 2008)21, as these are not

unambiguously control or treatment cities.

Table 5: Treatment and control characteristics

Number of cities

2005 highest-polluting

station avg.

Avg. number of exceedance

days

Cities violating PM10 standard

2005-06

No violation or AP 31 27.0 22.8 31

AP only 9 28.6 25.6 5

AP with LEZ before Oct. 2008 4 28.9 22.5 3

AP with LEZ starting 2009-10 10 29.2 27.8 5

AP with LEZ planned with no start date 26 29.1 32.2 18

5. Results

5.1 Effect of Non-Attainment Status

First we test the effect of being found in violation of the PM10 standard by comparing the

cities that developed an action plan without an LEZ only (APO cities) to the attainment cities

(C). Although action plans would have started being implemented as early as 2006, there is a lot

of variation in exactly when cities were able to enact the details of the plans, i.e. building ring

roads or promoting public transportation. To avoid this ambiguity, we compare PM10 levels in

20 Some cities preemptively implement an action plan to avoid violating the limits in the future, especially considering the tightened limits that go into effect January 1, 2010. 21 We drop the three cities that did not exceed the threshold until 2007 because those cities' action plans, although formally in effect, might not be yet fully implemented.

17

the period before being found in non-attainment, 2005, to 2008, when cities violating the

standard in 2005 or 2006 had two to three years to implement action plans22.

Table 6 below shows that the treatment effect is not significantly different from zero at both

traffic and background stations for all cities (columns 1-3). When only looking at cities over

pop. 100,000, adopting an action plan only is associated with a 4.5 percent decrease in PM10 at

traffic stations, significant at the five percent level (column 5), although the effects at

background stations and both stations together are insignificant (column 4 and 6).

Table 6: Effect of action plan only on log PM10, Jan-Oct 2005 vs. 2008 All cities All cities >100,000

All station types

Traffic stations

Background stations

All station types

Traffic stations

Background stations

(1) (2) (3) (4) (5) (6)

Treatment effect 0.000459 0.00336 -0.0229 -0.00658 -0.0447** 0.00521 [0.0123] [0.0147] [0.0180] [0.0175] [0.0208] [0.0214]

Observations 35097 18710 16387 19354 9036 10318 Adjusted R-squared 0.599 0.624 0.564 0.616 0.639 0.583

All regressions include year-month fixed effects, weather, holiday, station type and population covariates

*** p<0.01, ** p<0.05, * p<0.1; Robust standard errors clustered by city-week in brackets

Next Table 7 shows how PM10 changed for all AP cities (all APO cities and all LEZ cities)

compared to the control cities (C). Again, for the larger cities with a population above 100,000,

there is no significant change in PM10 except for a small increase among background stations

that is only significant at the ten percent level (columns 4-6). Among all cities, including the

smaller cities below 100,000 PM10, however, PM10 slightly increases by 2.8 percent for AP

cities (column 1), with most of this increase coming from a 3.0 percent increase at traffic stations

while there is no increase at background stations (columns 2 and 3).

In summary, despite the fact that overall PM10 levels decreases since 2006, there is little

evidence that implementing an action plan contributes to additional decreases in PM10. We now

turn to the evaluation of the LEZ program.

22 These regressions only include January through October since we do not have 2008 PM10 data in November and December.

18

Table 7: Effect of non-attainment status on log PM10, Jan-Oct 2005 vs. 2008 All cities All cities >100,000

All station types

Traffic stations

Background stations

All station types

Traffic stations

Background stations

(1) (2) (3) (4) (5) (6)

Treatment effect 0.0282** 0.0304** 0.00831 0.0206 -0.0204 0.0368*

[0.0117] [0.0143] [0.0168] [0.0168] [0.0204] [0.0204]

Observations 38613 20405 18208 22288 10149 12139

Adjusted R-squared 0.603 0.624 0.569 0.621 0.641 0.589 All regressions include year-month fixed effects, weather, holiday, station type and population covariates *** p<0.01, ** p<0.05, * p<0.1; Robust standard errors clustered by city-week in brackets

5.2 Effect of action plans and LEZs

In this section, we isolate the treatment effect of having LEZs as part of an action plan.

Table 8 below shows the results of comparing the four LEZs within our 25 to 33.5 range of 2005

PM10 levels that began before October 2008—Cologne, Mannheim, Reutlingen and Leonberg—

to the attainment control cities as above. Since there is a set date for the LEZ treatment to begin

in 2008, whereas the action plan treatment was spread out over 2006 to 2008, we use 2007 as the

baseline year. This allows us to avoid any other confounding events that might have affected

PM10 differently due to the city specific contents of the action plans between 2004 and 2006.

We use April through October data to estimate Equation (2), since the Mannheim, Reutlingen

and Leonberg LEZs didn’t take effect until March 2008 and we are allowing a one month lag for

the city to adjust to the policy23.

Table 8: LEZ vs. Attainment cities, April-October 2007 vs. 2008 (1) (2) (3) (4) (5) (6)

All cities Cities >100,000

All station types

Traffic stations

Background stations

All station types

Traffic stations

Background stations

Treatment effect -0.0169 -0.0647*** 0.0197 0.0184 -0.0482* 0.0611**

[0.0182] [0.0193] [0.0236] [0.0215] [0.0249] [0.0251]

Observations 16,240 7,252 8,988 8,561 3,425 5,136

Adjusted R-squared 0.595 0.636 0.543 0.621 0.617 0.581 All regressions include year-month fixed effects, weather, holiday, station type and population covariates *** p<0.01, ** p<0.05, * p<0.1; Robust standard errors clustered by city-week in brackets

23 In some of the early LEZs, like Berlin, in the first few weeks often drivers were only warnings and not tickets. not ticketed immediately, but first a warning was spelled out.

19

As Table 8 shows, these four LEZs have lowered PM10 levels at traffic stations. On average

there is a decrease of 6.5 percent among all traffic stations (column 2) and 4.8 percent among

traffic stations in cities over population 100,000 (column 4). The treatment effect at background

stations is opposite, however. In the bigger cities, there is actually a 6.1 percent increase in

PM10 after the LEZ takes effect (column 6). Thus while pollution may decrease on the major

roads, this decrease is not being realized outside of these high-traffic areas. In the regressions to

follow we find this result holds for most LEZ cities: they see an increase in PM10 at background

stations relative to their counterparts who have not initiated LEZs yet. This could be further

evidence that traffic is being diverted from downtown areas, actually increasing PM10 in the

outlying areas.

Table 9 shows the effect of each of the four LEZs separately. Each LEZ is compared to a

smaller three to four PM10 μg/m³ band of control cities based on the 2005 PM10 readings.24

Panel A shows that the Mannheim LEZ decreased PM10 considerably at traffic stations—

between 14 and 22 percent when looking at all control and the larger control city25,

respectively—while PM10 levels rose 10 percent at background stations. Panel B shows that

none of the treatment effects are significant for the Cologne LEZ, although the coefficients are

negative for the traffic stations, while mixed for all background stations versus background

stations in large cities. The Reutlingen LEZ, on the other hand, is actually associated with an

increase in PM10 according to Panel C, although the coefficients are only significant when using

large comparisons for background stations only (9.2 percent) and background and traffic stations

together (6.1 percent). Panel D shows that the Leonberg LEZ decreased PM10 at its only traffic

station between 5 and 7 percent depending on whether controlled to all cities or only cities under

population 100,000 (Leonberg is under pop. 100,000). Thus while it is clear there is some

heterogeneity if the effects of these LEZs, there is a trend that LEZs are associated with

decreases in PM10 primarily at traffic stations, while PM10 increases at background stations.

24 Colonge and Leonberg are compared to attainment cities with highest 2005 PM10 readings between 25 and 27.5; Reutlingen is compared to attainment cities with 2005 readings between 27.7 and 30.2; Mannheim is compared to attainment cities between 30 and 33.5. 25 Columns 1a-3a compare Mannheim to Münster and Fürth, while columns 4a-6a only compare to the Münster, which is closer in size and 2005 PM10 levels.

20

Table 9: Effect of individual LEZs on log PM10

All cities All cities >100,000

All station types

Traffic stations

Background stations

All station types

Traffic stations

Background stations

Mannheim LEZ (Apr-Oct)

(1a) (2a) (3a) (4a) (5a) (6a)

Treatment effect -0.0159 -0.144*** 0.109** -0.0272 -0.222*** 0.109**

[0.0394] [0.0367] [0.0535] [0.0442] [0.0404] [0.0535]

Observations 2542 1258 1284 2114 830 1284

Adjusted R-squared 0.639 0.679 0.563 0.641 0.677 0.563

Cologne LEZ (Feb-Oct)

(1b) (2b) (3b) (4b) (5b) (6b)

Treatment effect -0.0121 -0.0227 -0.00758 0.0152 -0.00744 0.0267

[0.0334] [0.0396] [0.0347] [0.0330] [0.0396] [0.0342]

Observations 12582 5471 7111 5471 1642 3829


Reutlingen LEZ (Apr-Oct)

(1c) (2c) (3c) (4c) (5c) (6c)

Treatment effect 0.0167 0.00819 0.0413 0.0607* 0.0493 0.0923**

[0.0319] [0.0381] [0.0450] [0.0331] [0.0401] [0.0421]

Observations 4307 1739 2568 2595 1311 1284


Leonberg LEZ (Apr-Oct)

All cities All cities <100,000

(1d) (2d) (4d) (5d)

Treatment effect -0.0567* -0.0567* -0.0738** -0.0619*

[0.0298] [0.0307] [0.0308] [0.0318]

Observations 8963 4255 5967 3399

Adjusted R-squared 0.584 0.618 0.596 0.627

All regressions include year-month fixed effects, weather, holiday, station type and population covariates

*** p<0.01, ** p<0.05, * p<0.1; Robust standard errors clustered by city-week in brackets

21

5.3 Individual LEZ results

As a robustness check of whether the preceding results hold, we more carefully analyze the

major LEZs by comparing them to nearby cities using the difference-in-differences approach

specified in Equation (2). There are multiple advantages of looking at each LEZ one-by-one.

First, weather and geography vary considerably across Germany and this approach allows us to

fit a separate weather model for each region. Second, given the differences in the size of the

LEZs, it is likely that there will be heterogeneity in their effectiveness. we make use of the

staggered system of the introduction of LEZs, by comparing current LEZ to future LEZ cities.

Hence we compare LEZ cities that violated the limits in 2006 and/or 2007 to cities that are also

in non-attainment, also violating the limits in 2006 and/or 2007 but decided--for one reason or

another--to set the introduction date of the LEZ at a later point. This procedure, however, comes

at the cost that we do not primarily match on 2005 PM10 levels, but we rather match on (a) the

fact that all cities consider to implement an LEZ, (b) geography and (c) city size.The

identification in this section comes from the assumption that there are no systematic differences

in changes in LEZ cities’ PM10 levels based on when they implemented their LEZ beyond the

effect of the LEZ.

In the plots below, for each LEZ city we present the DD treatment effect coefficients with 95

percent confidence intervals from comparing both the background and traffic stations to

neighboring future LEZ cities26. These control cities are ranked in terms of how similar they are

to the LEZ city in terms of location, population and pre-regulation PM10 levels, such that the top

cities in each plot are the ‘best’ controls for each LEZ. In addition, for the LEZ cities that have

stations both within and outside of the LEZ, we show how PM10 levels have changed within the

city, per Equation (3).

The following regional regressions are consistent with the above findings. LEZs decrease air

pollution in high-traffic areas, but outside of these areas emissions tend to increase.

26 Regressions control for year-month fixed effects, station fixed effects, day of week fixed effects, holidays, polynomial and lagged weather variables.

22

5.3.1 Baden-Württemberg

The largest concentration of early LEZs is in the state of Baden-Württemberg. In the greater

Stuttgart area, LEZs came into effect on March 1, 2008, in Stuttgart, Ludwigsburg, Tubingen,

Leonberg, Reutlingen, Ilsfeld and Pleidelsheim. For the Stuttgart LEZ—the largest city in

Baden-Württemberg and largest LEZ by area in Germany—the LEZ has had a negative effect on

PM10 levels at traffic stations when compared to traffic stations at all but the least appropriate

nearby future LEZ control city (Mühlacker). As shown below in Figure 5, except for one outlier,

Ulm, all the coefficients are close to the average treatment effect of a 6.0 percent decrease in

PM10. At background stations, however, all the coefficients are positive, meaning there has

been an increase in background levels of PM10 in Stuttgart relative to other non-LEZ cities.

Figure 5: Stuttgart LEZ

For the Tübingen, a city 44 kilometers southwest of Stuttgart, the story is very similar. As shown in

Figure 6, the treatment effect is negative for all nearby future LEZ cities except Mühlacker,

and all the coefficients besides Ulm are around the average 5.9 percent decrease in PM10.

Again, the coefficients on background stations are all positive.

Mühlacker

Herrenberg

Ulm

Pforzheim

Freiburg

Heilbronn

Heidelberg

Karlsruhe

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients

Traffic stations Background stations

Suttgart LEZ

23

In nearby Ludwigsburg (Figure 7), the effect of the LEZ is not as clear. With the exception

of Ulm, all the coefficients are close to zero, with most confidence intervals including positive

and negative numbers. This may be because although Ludwigsburg is the second largest city in

Baden-Württemberg, its LEZ is much smaller—30 km2 and 55,000 inhabitants, versus Stuttgart’s

207 km2 and 590,000 inhabitants.

Leonberg is a considerably smaller city in the greater-Stuttgart area. As shown in Figure 8

above, the results when comparing to other smaller future LEZ cities are mixed and generally not

significantly different from zero. We do not study the LEZs in Ilsfeld and Pleidelshiem because

these cities are very small.

Figure 6: Tübingen LEZ

Mühlacker

Herrenberg

Ulm

Pforzheim

Freiburg

Heilbronn

Heidelberg

Karlsruhe

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Tübingen LEZ

24

Outside of the greater-Stuttgart area, there is an LEZ in Mannheim, Baden-Württemberg’s

third-largest city. While this LEZ is relatively small at 7.5 km2, it covers 93,000 inhabitants.

Mühlacker

Herrenberg

Ulm

Pforzheim

Freiburg

Heilbronn

Heidelberg

Karlsruhe

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Ludwigsburg LEZ

Mühlacker

Herrenberg

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients

Traffic stations

Leonberg LEZ

Figure 7: Ludwigsburg LEZ

Figure 8: Leonberg LEZ

25

The treatment effect when comparing Mannheim (Figure 9) to nearby future LEZ cities

Heidelberg and Karlsruhe are negative for traffic stations (average -9.7 percent) and positive for

background stations (average 10.4 percent).

5.3.2 North Rhine-Westphalia

Cologne is the only LEZ in North Rhine-Westphalia that came into effect in January 2008.

After Cologne, the next four largest cities in the state—Dortmund, Essen, Düsseldorf and

Duisburg—all have LEZs planned (all starting October 2008 except Düsseldorf, which began

February 2009). As seen in Figure 10 below, the coefficients for the Cologne LEZ regressions

are rather different from those seen in Baden-Württemberg, with the traffic stations experiencing

more increases in PM10 after the LEZ took effect and the background stations experienced more

decreases in PM10.

Karlsruhe

Heidelberg

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Mannheim LEZ

Figure 9: Mannheim LEZ

26

5.3.3 Hannover

The LEZ in Hannover is mid-size, covering 50 km2 and 218,000 inhabitants. Similar to the

other LEZs, Figure 11 shows that all of the coefficients for traffic stations are negative (average

8.6 percent decrease), and all but one of the coefficients for background stations is positive

(average 6.6 percent increase).27

27 In this case we compare Hannover to two action plan cities without LEZs planned, Göttingen and Osnabruck, since there were not many nearby similar cities with LEZs planned.

Dusseldorf

Duisburg

Essen

Dortmund

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Cologne LEZ

Figure 10: Cologne LEZ

27

5.4 Berlin

Our last section analyzes Berlin, the only city for which we have background and traffic

stations both within the LEZ and outside of the LEZ. The Berlin LEZ is the largest by

inhabitants, covering 88 km2 and 1.1 million people. While the LEZ officially started January 1,

2008, police did not start ticketing cars until March 1, 2008, so we define the LEZ treatment as

starting in March and drop January and February since the effect of the LEZ will be ambiguous

in these months. Unlike most LEZ cities, we have both traffic and background stations within

and outside of the Berlin LEZ. The first panel of Figure 12 shows the coefficients when

including all Berlin stations in the city-by-city comparisons, while the second panel looks only at

stations within the LEZ. The treatment effect of the LEZ is mostly negative for traffic stations,

especially amongst the ‘better’ control cities. The coefficients are larger when looking at

stations within the LEZ only, with an average decrease in PM10 of 7.6 after introduction of the

LEZ, implying that the LEZ lowered PM10 more on roads within the LEZ that roads outside of

the LEZ. The story at background stations is less clear, with the treatment effect being positive

when comparing to some cities and negative when comparing to others.

Osnabruck

Braunschweig

Gottingen

Bremen

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Hannover

Figure 11: Hannover LEZ

28

To further explore this, Table 10 shows the results of estimating (3), where the stations

within the Berlin LEZ are compared to those stations within Berlin but outside of the LEZ. The

traffic stations within the LEZ experience a 6.7 percent decrease in PM10 relative to traffic

stations outside the LEZ, either because PM10 emissions are decreasing within the LEZ, or

emissions are increasing outside of the LEZ as cars are forced to drive around it. This decrease

is 7.7 percent during for weekdays, implying the LEZ is slightly more binding for weekday

Cottbus

Potsdam

Madgeburg

Halle Saale

Dresden

Leipzig

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Berlin LEZ: All Stations

Cottbus

Potsdam

Madgeburg

Halle Saale

Dresden

Leipzig

All cities

-.6 -.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5 .6Coefficients


Berlin LEZ: Stations in LEZ only

Figure 12: Berlin LEZ

29

commutes. Conversely, the background stations within the LEZ see an increase in PM10 after

the LEZ takes effect than the background stations outside the LEZ.

Table 10: Berlin LEZ: Stations within LEZ compared to those outside

(1) (2) (3) (4)


All days Mon-Fri All days Mon-Fri

LEZ treatment -0.0668*** -0.0768*** 0.0568*** 0.0585***

[0.0217] [0.0243] [0.0127] [0.0145]

Observations 1960 1400 2938 2098

Adjusted R-squared 0.628 0.632 0.675 0.69

Robust standard errors in brackets, *** p<0.01, ** p<0.05, * p<0.1

6. Spatial Substitution between low and high emission cars.

One important argument in favor of the four-tier PM10 categorization is that it leads to a

more rapid adoption of cleaner technologies since even vehicle owners who do not typically

drive into an LEZ may want to keep the option value of free passage. The next task of this paper

studies the adoption of lower emitting vehicles and changes in the spatial composition of the

vehicle fleet. Using a unique panel dataset that provides the emission category and registration

location of each privately and commercially owned German vehicle from 2006 to 2009, we

analyze the spatial substitution effects in purchasing new vehicles and retrofitting existing high

emission cars due to the LEZ regulation.

6.1 Data

Data were obtained by Federal Motor Transport Authority (Kraftfahrtsbundesamt Flensburg)

in Germany. Yearly observation of the total number of cars by emission type (red, yellow, green,

no sticker) for all German districts, recorded as of January 1st, are available for the period of

2006 to 2009. Also, we have information on whether the vehicle is privately owned or

commercial used, including all of the commercial trucks.

Various German district reforms led to changes in the geographical boundaries of the

districts between 2006 and 2009. To account for these changes, several original counties had to

30

be merged into larger geographical units. As a result, the 411 current German districts (as of

2009) are reduced in our final dataset to 405 “counties”.28 The details of this procedure are

outlined in Appendix E. Error! Reference source not found. summarizes the composition of

the German vehicle fleet and Error! Reference source not found. shows how it has changed

since 2006.

Table 11: Vehicle Registration by emissions sticker category private vehicles Private Vehicles

2006 2007 2008 2009% change 2007 vs. 06

% change 2009 vs. 08

Green 31,039,096 33,226,606 34,020,748 34,862,420 7.05 2.47

Yellow 7,839,624 6,428,598 3,931,262 3,597,594 ‐18.00 ‐8.49

Red 1,781,310 1,634,275 1,267,825 1,092,315 ‐8.25 ‐13.84

No Sticker 4,885,299 4,656,493 1,597,089 1,381,064 ‐4.68 ‐13.53

Total 45,545,329 45,945,972 40,816,924 40,933,393 0.88 0.29

Table 12: Vehicle Registration by emissions sticker category commercial vehicles Private Vehicles

2006 2007 2008 2009% change 2007 vs. 06

% change 2009 vs. 08

Green 1,005,633 897,017 609,948 792,577 ‐10.80 29.94

Yellow 1,045,213 1,132,907 945,181 844,803 8.39 ‐10.62

Red 28,408 163,166 469,853 413,133 474.37 ‐12.07

No Sticker 734,275 658,928 524,542 518,545 ‐10.26 ‐1.14

Total 2,813,529 2,852,018 2,549,524 2,569,058 1.37 0.77

28 Counties is written in parenthesis because these “counties” include larger geographical regions due to the various mergers.

31

Figure 13: Change of share of Green Sticker vehicles 2009 to 2008 as function of distance of the county to LEZ (privately owned cars)

In the following we analyze the composition of the fleet of privately owned vehicles in 2008

and 2009. Figure 13 shows that between 2009 and 2008 the difference in the share of Green

Sticker vehicles (the cleanest technology) is between 0.01 and 0.035 share points. Moreover, the

Figure shows that closer the counties are located to an LEZ, the starker the increase in green

stickers.

Visually, in the scatter plot of Figure 13, Regensburg and Bonn are outliers. It turns out

however, that these cities’ special circumstances can easily explain their greater adoption of

green sticker cars. In 2007, the local government of Regensburg announced the introduction of

an LEZ in spring 2008, then decided to postpone the introduction until September 1st of 2008.

However, this date was then again postponed for prospectively January of 2010 (Stadt

Regensburg, 2008). We see that with the announcements however, the inhabitants of Regensburg

already preemptively responded in upgrading their vehicles. The second outlier is Bonn. It is just

a very short drive to Cologne–over highway 555 or highway 59–with (partially) no speed limits.

This closeness of Bonn to Cologne provides an incentive to obtain a green sticker.

.01

.015

.02

.025

.03

.035

0 50 100 150 200 250NEAR_DIST__km_

95% CI Fitted valuesyPKWdiff08to09

Bonn

Regensburg

Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Chang of share of Green Sticker vehicles 2009 to 2008

Distance of county to next LEZ (km) Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Change of share of Green Sticker vehicles 2009 to 2008

32

Figure 14 shows that between 2009 and 2008, the difference of the share is negative of

private cars that have the dirtiest technology (No sticker) for all counties in Germany. Moreover,

the Figure shows that closer the counties are located to an LEZ, the starker the decrease in No

stickers.

Figure 15 and Figure 16 show the changes in shares of yellow and red sticker vehicles,

respectively. Here the changes are more uniformly across different counties. This is not too

surprising since these are middle categories, which are banned by few of the current LEZs.

In summary, we find evidence that the introduction of LEZs does create an incentive for

drivers to substitute towards lower-emitting vehicles. The closer a county is to an LEZ, the more

likely its citizens have been to substitute away from the dirtiest cars and towards the cleanest

cars.

Figure 14: Change of share of No Sticker vehicles 2009 to 2008 as function of distance of the county to LEZ (privately owned cars)

-.0

25-.

02

-.0

15-.

01

-.0

050

0 50 100 150 200 250NEAR_DIST__km_

95% CI Fitted valuesyPKWdiffNS08to09

Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Chang of share of “No Sticker” vehicles 2009 to 2008

Distance of county to next LEZ (km) Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Change of share of “No Sticker” vehicles 2009 to 2008

33

Figure 15: Change of share of Red Sticker vehicles 2009 to 2008 as function of distance of the county to LEZ (privately owned cars)

Figure 16: Change of share of Yellow Sticker vehicles 2009 to 2008 as function of distance of the county to LEZ (privately owned cars)

-.0

1-.

008

-.0

06-.

004

-.0

02

0 50 100 150 200 250NEAR_DIST__km_

95% CI Fitted valuesyPKWdiff_R08to09

-.0

2-.

015

-.0

1-.

005

0

0 50 100 150 200 250NEAR_DIST__km_

95% CI Fitted valuesyPKWdiff_Y08to09

Regensburg

Distance of county to next LEZ (km) Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Change of share of Red Sticker vehicles 2009 to 2008

Distance of county to next LEZ (km) Blue line: fitted values of quadratic regression Grey area: 95% confidence interval Green dot: Change of share of Yellow Sticker vehicles 2009 to 2008

34

7. Conclusions

The 2005 passage of the European Union Clean Air directive and the resulting

establishments of the Low Emission Zones (LEZ) marked an unprecedented attempt to mandate

lower levels of air pollution in order to protect the public health in the European Union. In the

case of particulate matter, thought to be the most malicious form of air pollution, the EU set

maximum allowable concentrations that every measurement station in the EU is required to

meet. Those cities exceeding the allowable concentration must implement clean air action plans,

which have led to a multiplicity of urban policy measures. Different forms of implemented

traffic restrictions range from increasing traffic toll payments (i.e. London), environmental zones

(i.e. Berlin, Munich) or temporal complete traffic restrictions (Milan). It is debated however,

whether these methods work efficiently and which one is the most appropriate. This is of urgent

importance as the PM10 regulations will be substantially tightened in 2010, and many more

cities will have to implement policies to decrease pollution levels. This paper quantitatively

assesses LEZs, a policy popular in Germany, in terms of (a) air pollution outcomes and (b) the

effects on the spatial substitution between high and low emission cars.

This paper seeks to contribute to the environmental evaluation literature by assessing the

consequences of the widespread adoption of LEZs across German cities. Our results imply that

success of LEZs in lowering PM10 levels is mixed. Some of the biggest LEZs, such as Stuttgart

and Berlin, appear to have lowered PM10 at traffic air quality monitors within the LEZ. For the

one LEZ where we can compare traffic stations both inside and outside of the LEZ, Berlin, there

is a difference in the effectiveness of the LEZ—either the PM10 is decreasing more in the LEZ

than in surrounding areas, or drivers are increasing emissions outside of the LEZ by driving

around it. Consistently with the last assertion, most LEZ cities have seen an increase in PM10 at

background stations relative to their counterparts who have not initiated LEZs yet. This could be

further evidence that traffic is being diverted from downtown areas, actually increasing PM10 in

the outlying areas. It could also imply that cities that implement LEZs later may spend more

time on other action plan items that decrease pollution across the city, not just in urban city

centers.

To summarize, we find that the there is some evidence that LEZs have been effective at

lowering PM10 levels by four to seven percent. Furthermore we find that developing an action

35

plan only, but not having an LEZ, so far has not led to significant reductions in PM10 levels.

Most cities’ LEZs will become more stringent in the coming years, however, so the ultimate

effect of LEZs is still an open question.

Literature

“22nd Ordinance for the Implementation of the Federal Immission Control Act (Ordinance on

Immissions of Airborne Pollutants – 22. BImSchV), of 11 September 2002, Federal Law

Gazette I 2002, p. 3626.

“Air pollution: Commission starts legal action against 10 Member States over airborne

particles”. Europa Press Release (2009) IP/09/174.

http://europa.eu/rapid/pressReleasesAction.do?reference=IP/09/174

"Beijing extends vehicle limits for another year," (2009). Beijing Traffic Management Bureau.

http://www.bjjtgl.gov.cn/publish/portal1/tab165/info11307.htm

Bertrand, Marianne, Esther Duflo and Sendhil Mullainathan (2004). “How Much Should We

Trust Differences-In-Differences Estimates?” Quarterly Journal of Economics 119: 249-

275.

Cohen AJ, Ross Anderson H, Ostro B, et al. The global burden of disease due to outdoor air

pollution. J Toxicol Environ Health A 2005; 68: 1301–7.

“Council Directive 1999/30/EC,” (1999). Official Journal of the European Communities

L 163/60.

Council Directive 96/62/EC (1996). Journal of the European Communities 296, 21.11. p. 55–

63.

Davis, Lucas (2008). “The Effect of Driving Restrictions on Air Quality in Mexico City.”

Journal of Political Economy 116(1): 38-81.

Dingding, Xin (2009). "Beijing car restrictions to continue," China Daily.

http://www.chinadaily.com.cn/bizchina/2009-04/06/content_7651378.htm.

EEA (2008): Annual European Community LRTAP Convention emission inventory report 1990–2006. European Environmental Agency (EEA) Technical report 7/2008.

Europa Press Release (2008): Environment: Commission requests information from Member States on PM10 pollution,. IP/08/1112..

36

Goddard, Haynes C. (1997). “Using Tradeable Permits to Achieve Sustainability in the World’s

Large Cities: Policy Design Issues and Efficiency Conditions for Controlling Vehicle

Emissions, Congestion and Urban Decentralization with an Application Mexico City.”

Environmental and Resource Economics 10 (1): 63–99.

Hahn, Jinyong, Petra Todd, and Wilbert Van der Klaauw. 2001. “Identification and Estimation of

Treatment Effects with a Regression Discontinutity Design.” Econometrica 69 (1): 201–9.

Imbens, Guido and Jeffrey Wooldridge (2008). "Recent Developments in the Econometrics of

Program Evaluation." IZA Discussion Paper No. 3640.

Johannsen (2009): Schulferien.org website maintained Moritz Johannsen, 16548 Glienicke,

Germany.

Klinger, Matthias and Elke Sahn (2008). "Prediction of PM10 Concentration on the Basis of

High Resolution Weather Forecasting." Meteorologische Zeitschrift 17(3): 263-272.

Kotchen, Matthew and Laura Grant (2007). “Does Daylight Saving Time Save Energy?

Evidence From A Natural Experiment In Indiana.” NBER Working Paper 14429.

Krzyzanowski, Michal (2005). “Health Effects of Travel-Related Air Pollution: Summary for

Policy Makers. World Health Organization ISBN 92-890-1375-3.

Levinson, Arik and Sudhir Shetty (1992). “Efficient Environmental Regulation: Case Studies of

Urban Air Pollution in Los Angeles, Mexico City, Cubatao and Ankara.” Policy Research

Working Paper no. 942, World Bank, Washington, D.C..

Meyer, Bruce (1995). “Natural and Quasi-Experiments in Economics.” Journal of Business &

Economic Statistics 13: 151-161.

Roson, Roberto and Kenneth Small, eds. (1998). Environment and Transport in Economic

Modeling. Kluwer Academic Publishers, Dordrecht, Boston and London.

Rost, J., T. Holst, E. Sähn, M. Klingner, K. Anke, D. Ahrens, H. Mayer, in print. Variability of

PM10 concentrations dependent on meteorological conditions. – International Journal of

Environment and Pollution 36:3-18.

Small, Kenneth and Camilla Kazimi (1995). “On the Costs of Air Pollution From Motor

Vehicles.” Journal of Transport Economics and Policy 29:7-23.

Stadt Regensburg (2008): Stadtradt Regensburg, Ausschuss fuer Stadtplanung, Umwelt, Verkehr

und Wohnungsfragen. Beschluss vom 27. Februar, 2008, Regensburg.

37

“Time Extensions for PM10, Nitrogen Dioxide and Benzene.” Europa Press Release (2009)

http://ec.europa.eu/environment/air/quality/legislation/time_extensions.htm

UBA (2008): Phone conversation with UBA expert on PM10 statistics.

U.S. EPA. Air Quality Criteria for Particulate Matter (Final Report, Oct 2004). U.S.

Environmental Protection Agency, Washington, DC, EPA 600/P-99/002aF-bF, 2004.

Vukan R. Vuchic (1999). Transportation for Livable Cities, CUPR Press.

Watkiss, P. S. Pye S, Mike Holland (2005). CAFE CBA: baseline analysis 2000 to 2020. Report

to the European Commission DG Environment, Brussels.

Welty, Leah J and Scott L. Zeger (2005). "Are the Acute Effects of Particulate Matter on

Mortality in the National Morbidity, Mortality, and Air Pollution Study the Result of

Inadequate Control for Weather and Season? A Sensitivity Analysis using Flexible

Distributed Lag Models." American Journal of Epidemiology 162(1): 80-87.

Weitzman, Martin (1974). "Prices vs. Quantities." Review of Economic Studies 45: 703-24.

APPENDIX A: Data

To calculate daily averages, we first simply linearly impute the missing hourly readings

throughout the day29. Once we have daily averages, we interpolate the missing daily averages

for the 1.4 percent of days with no readings.

To make sure our results are not driven by changes in monitoring station composition, we

restrict our analysis to PM10 readings for stations that have readings for all of the years included

in each analysis30.

29 Among stations reporting half-hourly or hourly data, less than seven percent of days are missing observations for some hours over 70 percent of these being three hours or less. This imputation is done through linear interpolation 30 We define a station as having complete data for 2007 if there is data for at least 340 of the 365 days of the year. Since we only have data through October of 2008, a station has complete data for 2008 if there is data for 280 of the 305 possible days.

38

Appendix B: Current and Future German LEZs

City Start date Excluded vehicles Size of LEZ Inhabitants

Dates of future restrictions (2nd

round, 3rd round)

Future excluded vehicles (2nd

round, 3rd round)

Study LEZs (LEZs beginning in 2008)

Berlin 1/1/2008 no sticker 88 sq. km 1.1 mill 1/1/10 red + yellow

Bochum 10/1/2008 no sticker 58.1 sq. km 150,000 end of 2010 red + yellow

Bottrop 10/1/2008 no sticker 50 sq. km n/a end of 2010 red + yellow

Dortmund 10/1/2008 no sticker 19.1 sq. km 587,137 1/1/11 red

Dortmund (Brackeler Strasse) 1/12/2008

no sticker + red < 0.1 sq. km 300 1/1/10 not yet planned

Duisburg 10/1/2008 no sticker 100 sq. km Unknown end of 2010 red + yellow

Essen 10/1/2008 no sticker 140 sq. km 14,00 1/1/11 red

Frankfurt 10/1/2008 no sticker 110 sq. km n/a 01/01/10, 01/01/12 red, yellow

Gelsenkirchen 10/1/2008 no sticker 20 sq. km n/a end of 2010 red + yellow

Hannover 1/1/2008 no sticker 50 sq. km 218,000 01/01/09, 01/01/10 red, yellow

Ilsfeld 3/1/2008 no sticker 2.5 sq. km 4,000 01/01/12, not yet

planned red, yellow

Köln 1/1/2008 no sticker 16 sq. km 130 01/01/10, not yet

planned red, yellow

Leonberg 3/1/2008 no sticker 30 sq. km 40,000 01/01/12, not

planned red, yellow

Ludwigsburg 3/1/2008 no sticker 30 sq. km 55,000 1/1/12 red

Mannheim 3/1/2008 no sticker 7.5 sq. km 93,900 01/01/12, not yet

planned red, yellow

München 10/1/2008 no sticker 44 sq. km 431,000 01/01/10, not

planned red, yellow

Oberhausen 10/1/2008 no sticker 23.8 sq. km 91,000 end of 2010 red + yellow

Pleidelsheim 7/1/2008 no sticker 7 sq. km 7,000 01/01/12, not yet

planned red, yellow

Recklinghausen 10/1/2008 n/a <20 sq. km n/a n/a n/a

Reutlingen 3/1/2008 no sticker <10 sq. km n/a 1/1/12 red

Schwäbisch Gmünd 3/1/2008 no sticker 5 sq. km 20,000 01/01/12, not yet

planned red, yellow

Stuttgart 3/1/2008 no sticker 207 sq. km 590,000 01/01/12, not yet

planned red, yellow

Tübingen 3/1/2008 no sticker n/a n/a 1/1/12 red

City Start date Excluded Size of LEZ Inhabitants Dates of future Future excluded

39

vehicles restrictions (2nd round, 3rd round)

vehicles (2nd round, 3rd round)

Future LEZs (2009 and beyond)

Augsborg open no sticker 5.2 sq. km n/a 1/1/10 red

Bremen 1/1/09 no sticker 7 sq. km 56,000 1/1/10 red

Dresden 1/1/10 no sticker 4.2 sq. km 6,500 Unexplained n/a

Düsseldorf 2/15/09 n/a 13.8 sq. km 36,500 1/1/11 n/a

Freiburg 1/1/2010 no sticker 28 sq. km 120 1/1/12 red + yellow

Heidelberg 1/1/10 no sticker 10.3 sq. km 170 1/1/12 red

Herrenberg 1/1/09 no sticker n/a 28,000 1/1/12 red

Karlsruhe 1/1/09 no sticker n/a n/a 1/1/12 red

Mühlacker 1/1/10 no sticker n/a n/a 2012 red

Mülheim 10/1/2008 no sticker n/a n/a End of 2010 red

Neu-Ulm 1/1/09 no sticker n/a n/a 1/1/12 red

Osnabrück 1/4/10 no sticker 14sq. km 7,000 1/4/11 red

Pfinztal 8/1/2008 no sticker 31sq. km 18,000 1/1/12 red

Pforzheim 1/1/09 no sticker n/a n/a 1/1/12 red

Regensburg n/a no sticker n/a n/a n/a n/a

Ruhrgebiet 10/1/2008 n/a n/a n/a 1/1/11 orange

Ulm 1/1/09 n/a n/a n/a n/a n/a

Weitere mögliche n/a n/a n/a n/a n/a n/a

40

Appendix C: Sample Cities

City

Avg 2005 reading at

highest polluting station

Violate PM10

limits in 2005-06

Exceedance days in 2005 Attainment status

LEZ start date

Total population

Mönchengladbach 25.0 0 24 Attainment 261,216 Schwedt 25.0 0 23 Attainment 37,001 Ulm 25.1 1 18 AP with 'late' LEZ not set 120,748 Schweinfurt 25.1 0 14 Attainment 54,097 Altenburg 25.2 0 27 Attainment 37,236 Coburg 25.4 0 15 Attainment 41,768 Aschaffenburg 25.6 0 12 Attainment 68,645 Wiesbaden 25.8 0 18 Attainment 275,085 Bernhausen 25.9 1 21 AP only 13,216 Bautzen 25.9 0 20 Attainment 148,945 Weiden i.d.OPf. 26.0 0 22 AP with 'late' LEZ not set 42,603 Kelheim 26.1 0 26 Attainment 113,100 Stralsund 26.2 0 22 Attainment 58,563 Heilbronn 26.2 1 22 AP with 'late' LEZ 1/1/2009 121,498 Lindau (Bodensee) 26.3 1 28 AP only not set 79,636 Emden 26.3 0 20 Attainment 51,666 Nauen 26.4 0 25 AP with 'late' LEZ not set 16,674 Hanau 26.5 0 20 Attainment 88,251 Weißenfels 26.6 0 32 Attainment 73,624 Pirmasens 26.6 0 16 Attainment 42,761 Bamberg 26.7 0 20 Attainment 69,746 Freiberg 26.7 0 33 Attainment 144,094 Leonberg 26.8 1 16 AP wtih LEZ 3/1/2008 45,537 Burghausen 26.8 1 27 AP with 'late' LEZ not set Stendal 26.9 0 18 Attainment 130,436 Saal a.d. Donau 26.9 0 25 Attainment 5,484 Sulzbach-Rosenberg 26.9 0 27 Attainment 20,409 Köln 27.0 0 14 AP wtih LEZ 1/1/2008 986,317 Gelsenkirchen 27.0 0 24 AP with 'late' LEZ 10/1/2008 267,418 Mülheim 27.0 0 21 AP with 'late' LEZ 10/1/2008 169,651 Zittau 27.0 0 31 Attainment 29,898 Arzberg 27.0 0 24 AP with 'late' LEZ not set 5,893 Bösel 27.1 0 25 Attainment 7,562 Itzehoe 27.1 0 21 AP with 'late' LEZ not set 33,800 Dessau 27.2 0 18 Attainment 77,914 Schwandorf 27.3 0 30 AP with 'late' LEZ not set 144,644 Worms 27.5 1 27 AP with 'late' LEZ not set 81,984 Würzburg 27.7 0 30 AP with 'late' LEZ not set 134,080 Glauchau 27.8 0 24 Attainment 25,760 Norderney 27.8 0 17 Attainment 5,986 Wuppertal 28.0 0 20 AP with 'late' LEZ 5/12/2009 358,813 Aachen 28.0 0 18 AP only 258,055 Plauen 28.1 1 33 AP only 68,614

41

City

Avg 2005 reading at

highest polluting station

Violate PM10

limits in 2005-06

Exceedance days in 2005

Attainment status

LEZ start date

Total population

Ingolstadt 28.2 1 35 AP with 'late' LEZ not set 121,687 Magdeburg 28.3 1 22 AP with 'late' LEZ not set 229,344 Erlangen 28.3 0 22 AP only 103,469 Bielefeld 28.4 0 19 Attainment 326,336 Gera 28.4 1 31 AP with 'late' LEZ not set 103,446 Reutlingen 28.5 1 17 AP wtih LEZ 3/1/2008 281,933 Saarbrücken 28.5 0 18 Attainment 340,702 Neuwied 28.5 0 23 AP with 'late' LEZ not set 184,849 Koblenz 28.6 0 20 Attainment 106,238 Ratzeburg 28.8 0 28 AP only 13,671 Krefeld 29.0 1 24 AP with 'late' LEZ not set 237,336 Datteln 29.0 0 30 Attainment 36,297 Borna 29.1 0 31 Attainment 22,561 Neu-Ulm 29.1 1 34 AP with 'late' LEZ 1/1/2009 163,477 Jena 29.6 1 29 AP with 'late' LEZ not set 102,291 Landshut 29.7 1 39 AP with 'late' LEZ not set 61,757 Nürnberg 29.7 0 33 AP with 'late' LEZ 1/1/2009 498,936 Weimar 29.8 1 35 AP with 'late' LEZ not set 64,541 Trier 29.9 0 26 AP with 'late' LEZ not set 100,198 Karlsruhe 29.9 1 22 AP with 'late' LEZ 1/1/2009 285,756 Bottrop 30.0 0 33 AP with 'late' LEZ 10/1/2008 119,195 Fürth 30.1 0 30 Attainment 113,596 Wetzlar 30.2 0 24 AP only 52,831 Ansbach 30.2 1 29 AP only 40,531 Brandenburg an der

Havel 30.7 1 53 AP with 'late' LEZ not set 73,886 Lutherstadt

Wittenberg 31.2 1 42 AP with 'late' LEZ not set 47,540

Regensburg 31.6 1 37 AP with 'late' LEZ not set

130,153 Ludwigshafen 31.7 1 37 AP with 'late' LEZ not set 163,536 Görlitz 31.8 1 42 AP with 'late' LEZ not set 57,418 Hagen 32.0 1 27 AP only 196,295 Halle/Saale 32.2 1 51 AP with 'late' LEZ not set 236,576 Kassel 32.2 1 48 AP with 'late' LEZ not set 193,842 Aschersleben 32.2 1 38 AP with 'late' LEZ not set 31,717 Freiburg 32.5 1 21 AP with 'late' LEZ 1/1/2010 216,448 Münster 32.5 0 33 Attainment 271,404 Frankfurt 32.5 1 48 AP with 'late' LEZ 10/1/2008 648,925 Mannheim 33.4 1 43 AP with LEZ 3/1/2008 307,847 NOTE: AP with 'late' LEZs are Action Plans only cities in our analyses since we have no data while their LEZs are in effect.

42

Appendix D: Spatial Substitution Data Irregularities

To merge the car registration data from 2006 to 2009, we had to take into account several

reforms of districts. Sachsen-Anhalt had a district reform in 200731 The first row in Figure E.1

shows the districts as of 2006, which merged into the districts of 2007, displayed in the second

row.

Figure E.1: Merged districts in Sachsen Anhalt in 2007:

However, as it was not possible to do so for LK Harz, Salzandkreis, LK Anhalt-Bitterfeld,

LK Wittenberg and Jerichower Land, we created two virtual districts corresponding to the union

of LK Harz and Salzandkreis in one hand, and LK Anhalt-Bitterfeld, LK Wittenberg and

Jerichower Land, in another hand.

Hence, we created a new artificial district, as displayed in Figure E.2. For the years 2006 and

2007, we added the number of vehicles from the former districts to get a virtual number of cars

for the artificial new district. In 2008 this artificial new district is simply the merger of LK Harz

and Salzlandkreis. This had to be done because it is not possible to distinguish from the data,

how the number of cars in the district Aschersleben-Strassburg are divided into LK Harz and

Salzlandkreis.

31 This reform is a law in Saxony‐Anhalt, Germany, which came into effect on July 1, 2007, which outlines a reform

of the districts of Saxony‐Anhalt. It reduces the districts from 21 to 11. Nine new districts are created by

amalgamating existing districts, while Altmarkkreis Salzwedel and Stendal as well as the urban districts of Halle and

Magdeburg are untouched.

43

Figure E.2: Merged new artificial district in Sachsen Anhalt (SAR1) in 2007:

Similarly we proceeded with the merged districts displayed in Figure E.3 below.

Figure E.3: Merged new artificial district in Sachsen Anhalt (SAR2) in 2007

Moreover, we dropped Dessau-Rosslau as we do not have data for Rosslau in 2006 and 2007.

Similarly, to match the districts between 2008 and 2009 another 26 districts had to be merged.

For 2006 and 2007, the number of cars in each emission category are recorded separately for

petrol and diesel. For the data beginning of 2008, however, the number of cars powered by gas

and ethanol were added to the category of the cars that are powered by petrol. Typically cars

powered by gas an ethanol account for a very small number of the cars only.

The data from 2006 to 2007 include vehicles that were under temporal suspension (i.e.

vacation vehicles). From 2008 onwards however, all suspended cars are no longer listed in the

dataset. Hence from 2007 to 2008 the total number of officially recorded cars in Germany

dropped by 10%. Unfortunately it is unknown which sticker vehicles dropped most, but likely

the dropped ones belong to the higher/older polluting vehicles. In the current draft regressions

and Figures we hence use the consistent 2008 and 2009 data only.

44

Appendix E: Stuttgart LEZ

Copyright: Landesvermessungsamt Baden-Württemberg, Bundesamt für Kartographie und Geodäsie 2003

The English term “Low Emission Zones” is commonly known in German as Umweltzone (Environmental Zone)

Cars, Air Pollution and Low Emission Zones in Germany

Documents

Transcript of Cars, Air Pollution and Low Emission Zones in Germany